The African e-Journals Project has digitized full text of articles of eleven social science and humanities journals.   This item is from the digital archive maintained by Michigan State University Library. Find more at: http://digital.lib.msu.edu/projects/africanjournals/

Available through a partnership with

Scroll down to read the article.

Article

Class, Distribution and Redistribution inPost-Apartheid South Africa

Jeremy Seekings andNicoli Nattrass

Soon after democratisation in 1994, the last vestiges of statutory racialdiscrimination were removed in South Africa. But the new governmentinherited a society where inequality could not be reduced to race alone andinequality persisted in the face of formal political equality. The best availabledata suggests that the Gini coefficient remained stable or even edged upduring the 1990s, as inter-racial inequality declined but intra-racial inequalityrose sharply.

Why has the demise of apartheid and the onset of democracy not beenaccompanied by a decline in inequality? As Aristotle noted two and a halfmillenia ago, democracy entails rule by the poor because the poor constitutethe majority. Why hasn't the formal equality of representative democracyinduced the ANC to introduce more effective pro-poor reforms? Regrettablythere are no careful studies of policy-making in the post-apartheid statecorresponding to (say) Posel's study of influx control under the apartheidstate (Posel 1991). In the absence of such studies, we have to resort to thebroad sweep of political economy, the strength of which is the linkage ofpolitics to economics and the weakness of which is its treatment of the stateas a 'blackbox'.

The literature on the political economy of post-apartheid South Africancan be divided into three main approaches. The first approach, representedmost influentially in the work of Bond (2000) and Marais (1998, 2001),analyses South Africa as a battleground between capital (especiallyinternational capital and its supposed agents, the IMF and World Bank) andthe largely undifferentiated South African masses. Whilst neo-Marxist in itscritique of capitalism, this approach lacks any empirical analysis of the classstructure. Its empirical content comprises primarily a critical analysis of the

TRANSFORMATION 50 (2002) ISSN 0258-7696

Jaiwny SaeMng* and Nicotl Nattras*

ANC's macro-economic, housing and urban infrastructural policies, andmost recently the government's commitment to privatisation. Its politicalanalysis does not extend far beyond the assertion that the ANC leadershiphas 'sold out'.

A different approach is evident in the work of scholars with close linksto the labourmovement, including Adler and Webster (1999,2000; also Gelband Webster 1996, Webster 2001) and Baskin (2000). These scholars haveto explain the continued participation of COSATU in an alliance with theANC. For them, post-apartheid policy reflects a class compromise, in whichpro-capitalist macro-economic and other policies are weighed up againstlabour's achievements in terms of labour legislation (especially the LabourRelations Act) and corporatist institutions (especially NEDLAC). Scholarsin this tradition tend to see society very much in terms of two classes thatcorrespond to the 'two nations' identified by President Mbeki: one rich andwhite, the other poor and black, both largely undifferentiated (Adler andO'Sullivanl996).

Our own work falls into a third approach in the literature on politicaleconomy. We recognise that capital retains enormous power, not becauseof the voting power of capitalists in elections but because of their importanceto the economy. In an era of globalisation, in which capital is more mobileand can exercise more readily the 'exit option' outof a country (or simply notenter in the first place), the power of capital relative to both other classes andnational governments is surely greater than before (see Nattrass 1999). But,we have argued, it is simplistic to view a society like South Africa simply interms of capital and labour. There remain important divisions and conflictsof interest within what other scholars refer to as the 'working class as awhole' (Nattrass and Seekings 1998c, Seekings2000). Moreover, social andeconomic change has eroded greatly the correlation between race and classin South Africa. High levels of inequality are increasingly based on intra-racial not inter-racial inequalities (Nattrass and Seekings 2001b).

The class structure of post-apartheid South Africa comprises three majorgroups of classes in South African society (see figure 1). At the top is anincreasingly multi-racial upper class or elite. In the middle lie workers in arange of classes: the 'semi-professional' class (teachers and nurses), the'intermediate' class (ie white-collar workers in public and private sectors)and most of the 'core' or urban industrial working class. At the bottom arethe marginalised sections of the working-class (including especially farmand domestic workers and their dependants) and households where no one

Class, Distribution and Redistribution in Post-Apartheid SA

is in employment (see further Seekings and Nattrass, forthcoming: chapters7 and 8).

Figure 1: The Post-Apartheid Class Structure

A/ \ * Upperclass

/ \ Sara-prof./ \ class

/ \ ^ Intermediate/ \ class

/ \ Corewoddng/ \ class

/ \ Pedytraders

/ V 1 Marginal/ \ working class

Underclass andother

Meanhouseholdincome(RAnonth1993)

7020

32642257

1187

1442

618

413

%ofallhouse-holds

12

519

19

5

12

28

%ofallincome

45

45

10

Notes: Data is from the 1993 SALDRU survey. The 'upper class'comprises households headed by people in managerial, technical orprofessional occupations, or with substantial income from assets orentrepreneurial activities; the 'semi-professional class' compriseshouseholds headed by teachers or nurses; the 'intermediate class'comprises households headed by routine white-collar, skilled orsupervisory workers; the 'core working-class' comprises householdsheaded by semi-skilled or unskilled workers outside of agriculture anddomestic work; the 'marginal working class' comprises farm anddomestic workers; the 'underclass and other' category compriseshouseholds with no members in employment and negligible income fromentrepreneurial activity or assets.

The post-apartheid 'distributional regime' has its roots in the distributionalregime under apartheid, which itself reflected the class interests of powerfulwhite constituencies (Nattrass and Seekings 1997). The basic framework ofapartheid labour market and welfare policies was designed to protect theinterests of sections of the white working class in the 1930s. Massiveinvestment in public education for white children in the 1950s and 1960s

Jeremy Seeking* and Nlcoli Nattrass

resulted in white workers securing the skills that enabled them, in the 1970sand 1980s, to commandhigh incomes in free labour markets, largely removingtheir dependence on direct state interventions (such as job reservationthrough the 'colour bar') and even on the institutional framework for wagedetermination. By this time, however, the state' s labour market and welfarepolicies had begun to serve the interests of the newly skilled African workingclass, which effectively opposed the dismantling of a range of pro-labourpolicies. Policies that had protected the incomes of white workers from the1920s until the 1970s now served to protect the incomes of semi-skilled andsome unskilled African workers. To the extent that this encouraged job-shedding and capital-intensive growth, these policies were at the expenseof other unskilled workers and the growing ranks of the unemployed.

Today, the organised African working class enjoys household incomesabove the median but below the mean in South Africa. Whilst the class isclearly 'exploited' in some senses, and receives incomes that are low bycomparison with South Africa's privileged elite, it also seeks to protect itssemi-privileged position. It resists reforms to labour market and otherpolicies that would steer the economy down a more labour-absorbing growthpath (see Nattrass 1999,2000a, 2000b, 2001), and resists any extension of thetax base that might transform them from net beneficiaries of fiscalredistribution (as they are at present) into net contributors.

Compared to capital and organised labour, the poor have little leverageover policy-making. They cannot threaten to withhold their capital or theirlabour. But they are numerous, giving them potential electoral strength.Elsewhere (Nattrass and Seekings 2001 a) we identify three reasons why thepoor use so feebly this potential electoral strength. First, the evidencesuggests - contra the conventional wisdom - that the poor have limitedexpectations of material gains in the short-term, because they are bothpatient and they recognise the constraints on the state (see further Charney1995, Nattrass and Seekings 1998b). Secondly, poor voters in most parts ofthe country do not have the option of switching their support to pro-poorparties because either there are no such parties in the area or, if there are, theylack credibility. Thirdly, and perhaps most important of all, some policies aremore opaque than others. Poor voters may be very sensitive to the deliveryof services (such as public education, health care and welfare) that affectthem directly. But it is not easy for voters to discern precisely the extent towhich the government is responsible for the shortage of employmentopportunities for the unemployed. Thus, whilst the poor express criticism of

Class, Distribution and Redistribution in Post-Apartheid SA

union action that is seen to impede job creation and may engage in directaction against immigrants who are seen to be taking scarce jobs, they maynot punish the government for the lack of job creation. In the long-term,electoral self-interest might push the ANC toward more egalitarian policies,but pressures are weak in the short-term.

The result of these political forces is a set of government policies thatoffers something to everyone in a giant but precarious political compromise.Those groups with an interest in free markets get conservative macro-economic policies. The industrial working class gets protective labourmarket legislation, a favourable deal from the budget (especially through thecontinued restriction of the tax base), and have (to date) forestalled majorprivatisation. Many public sector workers (including, especially, teachers)continue to enjoy favourable salaries. The poor get a certain amount ofredistribution through the budget (including especially an old-age pensionsystem whose coverage and generosity is unique in the developing world).Racial rhetoric (including the 'two nations' thesis) helps the ANC topreserve its multi-class support base. If government performance is poor,the ANC argues, this is due to the apartheid legacy and beyond the ANC'scontrol. We argue that the legacy of the past is heavy, but that thegovernment has chosen not to tackle it fully - because of the politicalpressures on the government itself.

In this paper we examine, first, the distributive consequences of post-apartheid South Africa's economic growth path and of more obviouslyredistributive policies. We then turn to evidence on processes of classformation and changing patterns of income distribution. Finally, we identifya series of policy reforms or strategic emphases that would have moreegalitarian effects in South Africa's labour-surplus economy.

The Growth Path and RedistributionUnemployment is a major determinant of poverty and inequality, and therewas clear evidence of high and rising unemployment in the 1990s (Nattrass2000b). At the same time, the structure of South African employment hasbecome more skilled (Bhorat and Hodge 1999). This is the result of twotrends: a general shift away from unskilled labour in all sectors and theespecially sharp decline in the labour-intensive mining sector. Employmentshedding off commercial farms contributed to the shift away from unskilledemployment in South Africa over the decade (Simbi and Aliber 2000:7).

Jeremy Soefcings and NkxHi Nattrass

Figure 2 illustrates the dramatic decline in employment in mining andagriculture across the 1990s - especially with regard to less-skilled workersin those sectors. Thus, even in the historically unskilled labour intensivemining sector, there has been a shift towards higher paid, better skilledworkers - but at the cost of employment overall.

Figure 2. The Decline in Employment in Mining and Agriculture(Sources: South African Reserve Bank Data and data extrapolated fromEdwards (2000))

-Miningemployment(1990=100)

—Agriculturalemployment

1 (1990=100)

• Least skilledmining andagriculturalemployment

Real earnings per worker increased over the period (see Figure 3), butespecially in the late 1990s - being driven mainly by increases in the publicsector. Rising average earnings are a function of two factors: trade unionpressure (especially in the public sector) and the changing skill compositionof the employed workforce. Given that unskilled workers are bearing thebrunt of retrenchments, one would have expected average earnings to haverisen (as better-paid, relatively skilled workers remained in employment).One would thus expect labour productivity to be rising as well. As shownin Figure 3, this indeed appears to be the case: labour productivity has risenas employment declined.

This outcome is consistent with the goals and policies of the government's'high productivity now' (HPN) growth strategy (Nattrass 2001). Notice thatthe growth in labour productivity exceeded that of average wages for mostof the period - thus facilitating a slight growth in the profit share. Thissuggests that the growthpath has been relatively kind to employed (especiallyskilled) labour and capital, and unkind to unskilled labour and the unemployed.

Class, Distribution and Redistribution in Post-Apartheid SA

Figure 3. Labour productivity, employment, average wages and theprofit share (indexes, 1990=100)

140- Total non-agriculturalemployment

•Labourproductivity

•Total realremunerationper worker

. index ofgross profitshare

This is not to suggest that the growth path has heen good for capitalaccumulation. To the contrary, economic growth has been lacklustre andpercapita incomes have fallen across the 1990s. But some economic activitieshave suffered less than others - and have even benefited from the policyinterventions and institutional forces acting on the economy to drive it 'upthe value chain'. In other words, the HPN growth strategy has paid dividendsfor some workers and some capitalists, but it has done nothing to improvethe labour absorption capacity of the economy. The gap between labourmarket insiders and outsiders (especially the unemployed) is thus likely toremain large (if not widen) and there appear to be few forces acting to narrowoverall inequality in South Africa.

Has redistribution through the budget ameliorated the inegalitarianimpact of the growth path? The scope for increased redistribution throughthe budget in South Africa after 1994 was constrained by three factors. Thefirst was the government's commitment to conservative macro-economicpolicies, which precluded large increases in overall government spendingwithout matching increases in taxation. Taxes rose for just about everybodyin the mid-1990s (Simkins and Woolard 2000), but this was relayed intoreduced budget deficits rather than increased spending. The second was thefact that the government inherited, in 1994, a budget that was alreadysignificantly redistributive. Thirdly, the government remained subject topolitical pressures that diverted it from the policy reforms (particularly in thelabour market) that would have reduced inequality most markedly (byprioritising job creation).

Jeremy Seekings and Ntedi Natttass

The development of the government's macro-economic strategy hasbeen well documented (see eg Marais 1998,2001, Michie and Padayachee1998, Habib andPadavachee 2000), but recent fiscal incidence analyses havenot been acknowledge in the political-economic literature. McGrath et al(1997) have shown how redistributive the budget was at the end of theapartheid period, reducing the Gini coefficient from about 0.7 (for gross ororiginal income) to about 0.6 (for post-tax/transfer income, taking intoaccount also the value of' in kind' benefits from public education and healthcare). Redistribution happened because the top quintile received muchfewer transfers or in kind benefits than they paid in tax, whilst everyone elsereceived more benefits than they paid taxes.

Van der Berg (2000, 2001) argues that the budget became still moreredistributive after 1994. Taking into account government spending onwelfare transfers, public education, public health, subsidies for housing andcapital expenditure on the provision of water, he estimates that spending onthe poorest 40 per cent of households (ie quintiles 1 and 2) rose by about5 0 per cent in real terms between 1993 and 1997. A small part of this was madepossible by reduced spending on the rich, in that spending per capita onthe top quintile actually declined. But the lion's share of extra spending onthe poor in the mid- 1990s came from increased and well targeted spendingby the government.

This targeted expenditure entailed not cash income, in the form ofgovernment welfare transfers, but rather benefits in kind — especially interms of public education. Van der Berg shows that the removal of indirectdiscrimination in teachers' salaries, together with the provision of some extrateachers and hence reduction in pupil-teacher ratios, entailed massiveincreases in spending in 'African' schools, ie schools with overwhelminglyAfrican students, especially in poor, rural areas. Under apartheid, teacherswere paid on different salary scales. In 1996-7 all teachers were moved ontoa single, consolidated salary scale, based on the scale of the former whiteeducation departments. Approximately 40 per cent of teachers were movedinto higher salary brackets, and average salaries rose by between 12 and 15per cent (RSA 2000a:4.5). This shift was probably driven by the need toderacialise salaries, especially given pressure from African teachers, ratherthan a concern with the poor. But if the value of public education is deemedequal to the cost of providing it, then the poor can be said to have benefitedsubstantially from this increase. Van der Berg (2000) calculates that the shareof total spending on public education accruing to the lowest income quintilerose from 23 per cent to 29 per cent between 1993 and 1997.

Class, Distribution and Redistribution in Post-Apartheid SA

Describing an increase in teachers' salaries as 'increased spending on thepoor' is somewhat misleading-unless, of course, the increased salaries areaccompanied by improved quality of teaching. However, teachers in schoolsin poor areas remain inadequately qualified, often poorly motivated, and intoo many cases simply incompetent. It is probably fair to conclude that, atleast in the short-term, the major beneficiaries of increased educationalspending were teachers (who are not relatively poor), not the studentssitting in their classes.

In other areas of public spending the gains to the poor were alsoambiguous. The health, housing and infrastructural budgets may havebecome better targeted on the poor, but it is unclear how much the pooractually benefited in terms of the quality of the services provided.

If the evidence on benefits 'in kind' is ambiguous, the evidence on cashbenefits actually undermines the government's claims. The single mostimportant instrument of direct redistribution through the budget is the oldage pension. Between 1993 and 2000 the real value of the old age pensiondeclined by a total of about 20 per cent. The fact that spending on old agepensions declined faster than total government spending indicates'that theelderly and their dependants bore the brunt of the adjustment in this regard.Similarly, the savagery of the 1998 reform of public financial support for low-income single parents suggests that the government is passing a significantportion of the burden of budgetary austerity onto the poor.

Cuts in welfare benefits have been defended on the basis that socialwelfare is to be 'developmental' rather than dispensing 'handouts' (seediscussion in Budlender 2000:125,130-3). But what this means in practice isunclear, and in the current context of rising unemployment and sluggishgrowth, such an approach imposes costs on the poor (at least in the short-term).

In essence, it is probably fair to conclude that six year's-into the post-apartheid era, the distributional regime remains much the same as ithad beenduring the late apartheid era. Growthpath policies and labour market policiesbuttress the earnings of the middle class (including in the public sector) andthe industrial working class, whilst eroding the demand for unskilled labourand reducing the prospects that the unemployed would secure employment.The welfare system provides generous assistance to old-age pensioners,limited assistance to single parents, and very heavily subsidised publicservices for children especially - but no assistance to the bulk of the poor,whose poverty was due to the lack of job opportunities.

Jeremy Seekings and NicoJi Nattrass

The divide that existed between white and African households under theearly apartheid period continues to shift, separating growing numbers ofbetter off African households from the African poor. The proportion ofAfrican households resorting to the private sector for the provision ofwelfare has grown steadily. By 1997 almost one-fifth of the population wascovered by private medical aid schemes, and coverage was closely relatedto income (South African Health Review 1997:82). Most formal sectoremployees—including the core working and intermediate classes — are nowcovered by private sector retirement funds. Growing numbers of Africanpeople in urban areas now send their children to schools that are private orsemi-private (in that, whilst in the public sector, they charge high fees toprovide a superior education). A multi-tier system of education, health andretirement pensions has emerged, with access to the higher tiers dependenton prior employment.

As was the case with the apartheid distributional regime, the publicwelfare system makes no provision for the many poor people who are not oldenough for the pension nor young enough to qualify for child support.Despite the limited reform of unemployment insurance (which extendedcoverage to workers in previously excluded sectors, such as domesticlabour, and provided higher proportional benefits for low-income workersthan high-income workers) there is still no provision for the long-termunemployed, nor for people who have never been employed.

Whereas the apartheid distributional regime was premised on fullemployment, the post-apartheid distributional regime operates in the contextof extremely high unemployment (see Nattrass and Seekings 1997). Theabsence of any welfare net for the unemployed thus constitutes a majorproblem. The semi-privileged position of politically powerful African groups- including the urban, industrial working class, sections of the intermediateclass and the semi-professional class (especially teachers) - gives themgood reasons to oppose a universal welfare system in that radical welfarereform would require increased taxation on them, making them subsidisersof the poor rather than the beneficiaries of redistribution from the rich.Perversely, therefore, the legacy of apartheid included the formation of apolitically powerful, cross-racial coalition of classes, some a lot moreprivileged than others, with an interest in opposing radical reforms to thedistributional regime.

10

Class, Distribution and Redistribution in Post-Apartheid SA

Class Formation and InequalityThere is some evidence to suggest that inequality has grown in the 1990s.Whiteford and van Seventer (2000), using census data, show how thedistribution of income became more unequal between 1991 and 1996. Inter-racial inequalities declined (although they remained large, of course), whilstintra-racial inequalities continued to grow. The incomes of the richest ten percent of African households rose by 17 per cent, whilst the incomes of thepoorest 40 per cent of African households fell by 21 per cent. Rich whitehouseholds maintained stable incomes, but relatively poor white householdssaw a big absolute decline in their incomes (Whiteford and van Seventer2000:14-19).

The real winners in terms of rising incomes were the better off Africanhouseholds. Of the total real increase in income between 1991 and 1996,40per cent went to the richest ten per cent of African people, and a total of 62,5per cent went to the richest 40 per cent of African people. The poor majorityof African people barely benefited at all (Whiteford and van Seventer2000:1). This data can also be presented in terms of income categories. Theshare of total, ie national, income earned by African households with anincome of at least R72 000 (in 1996 prices) rose fromnine per cent in 1991 to14,5 per cent in 1996 (Whiteford and van Seventer 2000:22-4).

These findings are broadly corroborated by data from the KwaZulu-NatalIncome Dynamics Study (KIDS). KIDS found that the proportion of its panelwith incomes below a fixed poverty line rose from 35 per cent in 1993 to42per cent in 1998. At the same time, the proportion with high incomes alsogrew, indicating deepening polarisation (Carter and May 1999).

These trends are consistent with our argument about the post-apartheidgrowth path: employed workers, particularly those with skills, are relativelyadvantaged by HPN policies; whereas less-skilled workers and theunemployed are disadvantaged (at least in the short- to medium-term). Thegap between skilled and unskilled workers is thus likely to grow. Cichello etal' s (2000) analysis of changes in individual earnings in the KIDS data tendsto support this. They show that the earnings of workers in regular employmentrose by 37 per cent between 1993 and 1998, compared to an overall averagechange in earnings of just sevenper cent. The earnings of workers in regularemployment grew faster than the average for everyone in the sample. Someof this spectacular increase was because new entrants into formal employmenthad higher wages than those who left. But even among workers who werein formal employment in both 1993 and 1998, earnings rose by 20 per cent.

11

Jeremy Seekings and Nicoli Mattrass

Cichcllo ct al do not disaggregate the category 'regular employment' intodiscrete classes, but it is likely that the core working class and intermediateclasses enjoyed positive real income growth during the late 1990s. This doesnot mean - as Cichello et al emphasise - that all members of these classesprospered. The earnings of workers who lost or left their jobs plummeted.The aggregate gains of classes such as the core working class and intermediateclass should not obscure the fact that the composition of these classesshifted, as individuals and households dropped into lower classes, to bereplaced by individuals and households that were upwardly mobile throughthe class structure.

The Changing 'Middle Class'Has this slight widening of income inequality been accompanied by anymajor change in the class structure? Perhaps the most striking change inSouth African society in the 1990s has been the accelerated growth of whatis generally called the black or African middle class. Whiteford and vanSeventer's data shows that the lion's share of increased national incomebetween 1991 and 1996 accrued to a small but rich minority of African people.The combination of improved access to education, the removal of anyremaining restrictions on upward occupational mobility and the introductionof affirmative action policies by both the state and private sector meantrapidly expanding opportunities for some. This trend has presumably beenexacerbated by the flight abroad of many skilled white workers, intensifyingskills shortages.

Under apartheid, as Crankshaw (1997) has shown, upward mobilityamong African people was largely limited to the semi-professionaloccupations of teaching and nursing and white-collar occupations entailinglittle authority. Managerial posts were largely limited to the bantustanbureaucracies, and there were very few African businessmen. The end ofapartheid saw a major shift in the range of opportunities open to Africanpeople. The number of African employees in senior managerial positions inthe public and parastatal sectors grew rapidly. African managers were alsoappointed to senior managerial positions in the private sector, especially injobs where employers considered it important to have someone with a blackface or good political connections. As universities turned out more and moreblack graduates, so the proportion of professional and managerial postsfilled by African men and women rose.

Most remarkably, there was a clear increase in the number and importance

12

Class, Distribution and Redistribution in Post-Apartheid SA

of African businessmen. According to calculations by BusinessMap(1999:11,26-7), just one per cent of the market capitalisation on theJohannesburg Stock Exchange (JSE) was under black control in late 1995, butby January 1999 the share had risen to 5.5 per cent. This share was controlledby 3 5 listed companies. A further 11 per cent of JSE market capitalisation wasdescribed as being 'under black influence' (ie with black firms having asignificant but not controlling interest); by far the largest of these was SouthAfrican Breweries. Even in the aftermath of the 1998 stock market crash, thisamounted to almost R60 billion. Lacking capital, emerging African capitalistswere dependent on the financial institutions to buy into corporate ownership.The growth of the black corporate class in South Africa was thus based onthe country's well-organised banking sector, active equity market and thefact that control and ownership were already largely local rather thanmultinational — factors that distinguished South Africa from most otherAfrican countries (Randall 1996). But it has meant that 'black economicempowerment' has been driven by debt (with consequences for risk andhence entrepreneurial decision-making).

Black empowerment deals were concentrated in financial services (typicallywith black-owned firms forming partnerships with established institutions),information technology and telecommunications, and the media (includingpublishing). Penetration was more limited in the traditional industrial sectors,excepting the case of Johnnie. Whilst these emergent African capitalistswere not directly dependent on the state, they were especially active insectors or industries in which the state wields some control, for examplethrough licensing (eg fishing and media) or influence, through tendering,procurement or privatisation policies. Some of the fastest-growing 'black'investors were the ten or so trade union investment companies. Becausethey are not listed on the JSE it is apparently not possible to provide a clearaggregate figure for their investments (BusinessMap 1999:65).

There are not, of course, a huge number of African businessmen (orwomen). In no capitalist society do capitalists constitute more than a tinyproportion of the population. In terms of sheer numbers of people, it is theteachers, nurses, professionals and junior and middle managers whoconstitute the 'middle class'. In South Africa, there are still more whitepeople than African people in the middle classes (although it is likely thatwhite people comprised less than half of the top income quintile by the endof the 1990s).

The rapid growth of the black middle class has been noted by the

13

Jeremy Seekings and Nicoli Nattrass

marketing and market research industries. In 1998 the Bureau of MarketResearch at UNIS A conducted a survey on the lifestyle, beliefs and attitudesof the emerging African 'middle class' in Gauteng (defining households as'middle class', somewhat idiosyncratically, if they 'had moved to a betterdwelling or if there was an improvement in the occupation/employment ofan adult household member over the past five years' (BMR 1999:1). Manyhad moved into formerly white middle class residential areas. Most wereyoung, and had not iived in their present house for long.

This class had quite distinctive consumer preferences. The productwhich was seen as conferring the highest status on people was the car—withmen preferring BMWs or Mercedes, whilst women favoured VW Golfs orJettas. Cellphones and clothing also conferred status. They preferred to buyclothes from Foschini, rating especially highly Polo shirts and Pierre Cardinsuits. They aspired to shop at Woolworths. Over the past few years they hadcome to use financial institutions far more. The Standard Bank was thefavourite bank, and Old Mutual and Mutual and Federal the favouriteproviders of long- and short-term insurance respectively. They took theirholidays at inland resorts or, increasingly, overseas. A high proportion -almost 30 per cent - aspired to have their own businesses. Whilst the BMRinvestigation tends to treat the African middle class as a somewhat exoticspecies, it helps to fill the gap caused by the absence of detailed sociologicalor ethnographic studies of this new social group.

Survey and other data is beginning to suggest that this class willreproduce itself over time, ie that the privileges ofthe current generation willbe passed onto their children. There is a clear relationship betweeneducational attainment of children and the class ofthe household as a whole,even when inter-racial differences were removed from the picture. Figure 4shows the mean grade attainment of African children according to some ofthe classes defined as in Figure 1.

The African middle class is clearly growing rapidly, but is there anyevidence that the class structure as a whole is changing? Is it perhaps ratherthat the racial composition of different classes is changing? Whiteford andvan Seventer's calculations show that the number of households in the'middle class', defined simply in terms of income as households with anincome(in 1996prices) ofover R72000peryear,grewbetwecn 1991 and 1996,but not as fast as the population as a whole. The middle class thereforeactually shrank in relative terms. The African middle class grew rapidly,whilst the white middle class shrank very slightly. Such trends are consistent

14

Class, Distribution and Redistribution in Post-Apartheid SA

with the hypothesis that the overall class structure changed little, althoughthe racial composition of the upper income group was shifting rapidly.

highest school grade completed, bycurrentag* and dais, African children only, 1983

SPC

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Age

Note: data from 1993 S ALDRU survey. UC is upper class; SPC is semi-professional class; IC is intermediate class; CWC is core working classand MWC is marginal working class; data for other classes is notreported here.

Further research is required into the changing nature of the class structure,using data from surveys conducted after 1993. But the available data onincomes suggests that the biggest 'winners', as Whiteford and van Seventerput it, have been better-off African people who have moved into and through

15

Jeremy Ssekings and Nicdi Nattress

the 'middle class'. The losers, of course, have been those people withoutregular employment at all, many of whom do nothave the skills to secure jobsin a skill-intensive economy even if the growth rate picked up remarkably.Class inequalities show little sign of abating, even if the upward mobility ofaminority of African people means that inter-racial inequalities have droppedmarkedly.

The 'Re-segmentation' of the 'Working Class'A number of scholars, together with the trade union movement, point to newprocesses of segmentation within the labour market, between those inregular forms of employment and those in non-regular forms such as part-time, temporary and sub-contracted (out-sourced) employment. Elsewherewe argue that employers' recruitment strategies favoured some job-seekersover others, and suggested a large percentage of the unemployed should beconsidered as members of an 'underclass' on the grounds of thedisadvantages they suffer in terms of access to employment options (Seekingsand Nattrass forthcoming:chapter 8). Should we reassess our classcategories, and in particular the boundaries between core working class,marginal working class and underclass, in light of the shift from regular tonon-regular forms of employment?

Kenny and Webster, in an analysis of the growth of casualisation andsub-contracting, conclude that:

... profound changes are taking place in the labour market: new winnersand new losers are emerging. Above all, the resegmentation of the labourmarket is entrenching the position of the 'old losers' — the unemployed,informal sector, and migrant workers. Flexibilization is creating aninsecure, lower paid, and unprotected workforce. (1998:240)

Kenny and Webster attribute the shift to non-regular forms of employmentto employers' search for lower labour costs in the context of globalisationand 'neo-liberal' government policy. They pay little attention to the role oflabour legislation and other government policies that cannot be describedas typically 'neo-liberal'. As we have argued elsewhere, various governmentlabour market policies have the direct or indirect effect of pushing up therelative cost of unskilled labour (see eg Nattrass 2000a). It is not surprisingthat the shift to non-regular forms of employment has been especially markedin labour-intensive sectors, including cleaning, security, and the retailsector. Kenny and Webster acknowledge, but barely explore, the role ofunions in accentuating divisions in the workplace, as they protect theinterests of permanent workers against those of casual workers.

18

Class, Distribution and Redistribution in Post-Apartheid SA

Theron (2000) argues that we need to distinguish between casualisationand out-sourcing as the latter need not entail the former. However, inpractice, sub-contracting often entails a decline in hourly wages, benefitsand employment security. Some of the workers who are shifted from regularto non-regular forms of employment have skills and may realise steady oreven increased earnings. But it is likely that the bulk of newly casualised jobsinvolve the less skilled in occupations that we classified as core workingclass.

Does casualisation or out-sourcing change their class position? There isan argument that such workers should be classified in the marginal workingclass. In other words, changes in the labour market are such that marginalityshould not be defined in terms of sector only (with agriculture and domesticwork qualifying as marginal), but also in terms of the form of employment.If we were to do this, then it is likely that the marginal working class isgrowing and the core working class is shrinking.

Whether or not we interpret this shift as a changing class position, theextent of casualisation, out-sourcing and sub-contracting is unlikely to belarge relative to the very high and apparently, rising rate of unemploymentin South Africa. The marginalised sections of society are dominated by thelong-term unemployed, especially those in the 'underclass' with pooraccess to vacancies and hence little chance of securing employment.

Towards a Social Democratic AgendaIs there anything that government can do to reduce inequality whilstensuring a sustainable growth in income? In this concluding section wesketch the key components of what we consider to be a social democraticpolicy agenda for a middle-income labour-surplus economy like SouthAfrica's. We describe the agenda as social democratic because it combinesthe goals, of improved welfare (in the broad sense of the standard of livingof citizens), reduced inequality and extended democracy, to be achievedthrough a combination of state and market.

(a) Full employmentThe most important element of a social democratic agenda in a massivelylabour-surplus economy must be the reduction of unemployment throughsustained job creation. Sustained job creation requires mat the SouthAfrican economy achieves a higher rate of growth and follows a labour-absorbing growth path. Public policies must work towards both of thesegoals. We suspect that the state could pursue a less contractionary - or more

17

Jeremy Seeking and Nicdi Nattrass

expansionary - fiscal policy than it has under GEAR, and as critics in thelabourmovement and elsewhere have demanded. At the same time, however,it is imperative that any labour market policies that impede low-wage jobcreation be reformed (see eg Nattrass 2000a, 2000b). The poor cannot waitfor the benefits of a 'high-productivity now' growthpathto trickle down tothem. This probably means tolerating greater wage inequality (as low-wage,labour-intensive employment expands), in order to reduce overall incomeinequalities by creating jobs for the unemployed. We propose the reform oflegislation providing for the extension of centrally-bargained agreements,.increased scope for firm-level rather than industry-level bargaining, and thereform of the Employment Conditions Commission to ensure that account istaken of the employment effects of wage-setting. On this, our views contraststarkly with the official views of the labour movement.

We are not proposing the abandonment of employment protection orhealth and safety provisions for some workers and nor are we proposingwage cuts. Rather, entrepreneurs shouldbe allowed to create new employmentat low wages if necessary. Average wages will probably fall, but as a resultof job creation (which reduces inequality) not wage cuts (that probablyincrease it). This will help address the unemployment problem at the marginby nudging the economy in a more labour-intensive direction. It is onlythrough rapid growth that unempleyment will be significantly reduced.

Our position is a democratic one. Democracy and organisation should beextended into low-wage workplaces, not agreements bargained by other,self-interested parties. The extension of democracy into the workplaceshould surely entail allowing the people most affected by the consequencesof a decision to play the leading role in making that decision. Wageagreements should, as far as possible, be made by those workers who standto reap the benefits (higher wages) or suffer the consequences(unemployment) of the deal. In sectors where there is no collective bargaining,the Employment Conditions Commissions should be tasked with balancingthe need for acceptable minimum wages with job creation for the unskilledand unemployed.

The idea of employment subsidies has been raised, most notably by SamBowles (see eg Heintz and Bowles 1996) and introduced in a limited way inthe 2001 budget. The idea is that employers will be paid a wage subsidy whichwilllower the cost of employment without reducing workers'take-home payor aggregate demand in the economy. But the potential benefits need to beweighed up against the possible costs, which include subsidising inefficient

18

Class, Distribution and Redistribution In Post-Apartheid SA

use of labour, administration costs, deadweight losses (ie subsidisingemployment which would have been created without the subsidy) and so on(see Standing etal 1996:458-60). Taxing capital assets to pay for the subsidy(as suggested by Heintz and Bowles) should encourage labour-intensiveproduction at the level of the firm, but the economy-wide effects are unclear.The overall impact of financing the wage subsidy out of general taxation (asoperationalised by the government) is even more unclear. More research isrequired on this issue.

In an earlier debate, Gelb and Webster (1996) seemed to argue that theincomes ofthe 'working class as a whole' would be maximised through wageincreases, even if there is a trade-off between employment and wages. Thisargument makes no allowance for distribution within the 'working-class' -by which Gelb and Webster clearly meant all ofthe non-capitalist or non-bourgeois classes in society. As we argue repeatedly, reduced inequality,even within this 'working class as a whole', requires job creation. Gelb andWebster appear to hold to a utilitarian view of social justice: the wage billshould be maximised regardless of distribution. Our approach is more of aRawlsian one, in that priority should be given to improving the incomes ofthe poorest members of society. At the very least, the wage bill should notgrow in ways that benefit the semi-privileged (ie people with jobs) at theexpense of the poor.

In the short-term, the government might commit more funds to publicworks programmes. These are not as politically difficult as reforming labourmarket policies, as COSATU has already accepted the principle that publicworks programmes can pay low wages as a way of targeting the very poorand destitute. The Working for Water Public Works Programme is widelyregarded as a successful experiment, and could be expanded. Certainly, if thepolitical will was there, resources could be transferred from unproductivespending (as is the case withmuch ofthe defence budget) to such programmes— potentially with a dramatic impact on employment. The problem is thatunemployment in South Africa neither low nor temporary. In 1977, Chileemployed 5.5 per cent of the labour force in emergency public worksprogrammes, at wages of one third ofthe minimum wage (Cortazar 1997), butunemployment there was largely due to structural adjustment, and haddeclined to a very low level within a decade. In South Africa, public worksprogrammes would probably have to continue for a much longer period.

It is remarkable that the South African government has been so hesitantover the employment issue when it is such a pressing public priority. Opinion

10

Jeremy Seekings and Ntcoli Nattrass

polls routinely show that job creation is the issue regarded by most votersas the most important problem facing the country. Yet the government seemsalmost incapable of action. Indeed, there is no single member of the cabinetwithresponsibihty foremployment creation. Voters recognise this, dismissinggovernment performance on the issue as unsatisfactory or veryunsatisfactory. We hope that the democratic process will lead to heightenedpressure on the government to respond to public opinion.

(b) Education and Equal OpportunitiesInequality in South Africa is closely related to education. On the one hand,tertiary education opens up the route to 'middle class' occupations. On theother, young men and women who leave school prior to matric are usuallyconfined to a lifetime of low-paid employment or chronic or intermittentunemployment. How a student performs in school is, however, not simplya matter of individual ability. Rather, there is a clear relationship betweenparental class and children's educational attainment. A social democraticagenda must include the promotion of more equal opportunities. Underapartheid, massive investment in public education meant that the childrenof poor, mostly Afrikaans-speaking white workers did not end up with thesame lack of skills as their parents (and were thus able to secure better-paidemployment without explicit racial discrimination in the labour market). Thechallenge facing post-apartheid South Africa is how to promote betteropportunities for children from poor backgrounds, regardless of race, butwithin die budgetary constraints set by the available tax base.

Inequality in South Africa's schools is not primarily due to low enrolmentrates. South Africa has high enrolment rates for all income deciles andregardless of gender (Lam 1999). Nor, or at least less and less, is it due todiscrimination in spending: a larger share of public spending on educationis spent on the poorest quintile than the richest quintile, whilst differencesin spending per pupil are largely due to the fact that more of the children inthe richer deciles are in secondary schools (which are more expensive tooperate than primary schools). Rather, inequalities in education are due todifferences in the quality of schooling,, as well as factors relating to the familybackground of pupils. The Department of Education has itself identifiedcauses of poor performance: the poor educational background of parents,poor conditions of teaching and learning, inappropriate teaching andlearning methods, lack of access to reading and other educational materialsand libraries, poor school management, a lack of order and discipline among

20

Class, Distribution and Redistribution in Post-Apartheid SA

both teachers and pupils (which often results in the loss of time for teachingand learning) and the low morale of principals and teachers (RSA 2000:40-44). It is important to note that this list does not include pupil/teacher ratios,perse. Case andDeaton (1998), using the scanty statistics available, haveshown that pupil/teacher ratios are not a major factor in educational outcomes.

The Departments of Education and Finance, together with theParliamentary Standing Committee for Public Accounts, have begun to payattention to the problem of getting better' value-for-money' out of educationalexpenditure (see Seekings2001). School-children must achieve higher levelsof numeracy, literacy and life-skills for the same level of public expenditure.Research into this must pay careful attention to distributional issues, in thatthe goal of 'value-for-money' treats as valuable the achievement ofsubstantially equal opportunity for children from whatever background.

(c) Democracy-deepening asset redistributionExploitation of waged employees by profit-seeking employers may not be

' the only source of-injustice in society, but it is clearly an important one. Asocial democrat needs to balance a concern with justice in the employmentrelationship with the need to keep firms operating so as to provide a measureof employment security for workers — whilst providing workers with a realsay in the economic life and decisions of the firm. One way of doing this isto promote worker-ownership.

The government might provide developmental support for worker-ownedfirms. Thus, for example, if a clothing firm decided that it could no longercompete profitably on international markets, the government should consideroffering the workers (who face retrenchment) the option of taking over thefirm and working for a share of profits rather than the wage. This wouldenable them to earn less than the going wage - as they would be drawinga share of profits rather than the old wage which had put the firm underpressure in the first place. Such an arrangement would at least keep peopleemployed, albeit at lower earnings. However, there are reasons to believethat worker-owned firms can be more efficient than capitalist firms (Bowlesand Gintis 1998). Thus, with government support (perhaps in the form oftraining, such as community service for MBA graduates), it is possible thatworker-owners could end up earning more than they had as workers. Butwhatever the outcome, such creative solutions should be considered.

The idea of promoting worker-ownership in manufacturing and servicesis hardly on the public agenda at present, and would probably provoke much

21

Jeremy Sseldngs and Nicoli Nattrass

knee-jerk opposition. Yet there is widespread public support for workerownership in another sector of the economy: agriculture. Economic argumentsfor land reform typically combine concerns with efficiency and distribution:not only will income be more equitably distributed, in mat there is no divisionbetween profits to a few and wages to the many, but also there will be moreresources for the poor because production on small farms will be morelabour-intensive than production on large farms. Implicit in this argument arethe views that small farms will save on supervision costs (vis Bowles andGintis) or will work at lower marginal returns than the prevailing wage foragricultural labour. There are, of course, other arguments for land reform: itprovides security of tenure in accommodation, perhaps has social or culturalsignificance, and certainly has symbolic political importance ('Mayibuye iAfrica'!). But government grants for land reform are certainly motivated inpart by economic considerations.

Since 1994 the government has provided subsidies for land reform, toenable African small farmers to buy land at market prices from whitelandowners. The government initially itself the target of transferring about30 million hectares, or 30 per cent of the country's medium to high-qualityfarmland, to 600 000 households. To achieve this it provided a means-testedsubsidy equal in value to the subsidy available for housing. Small farmerswbuld have to pool their subsidies and whatever other capital they could laytheir hands on, and buy a farm from a landowner. It was not long before theMinister of Land Affairs decided that the plan was neither fiscally noradministratively viable. After three years, only 200 000 hectares had beentransferred, to about 20 000 poor households (Deininger and May 2000:8).Deininger and May, in their studjrof the land reform experiment, concludedthat, with minor reforms, it could make a major, contribution to reducingpoverty in rural areas. But in 2000, the new Minister of Land Affairs andAgriculture announced a major shift in policy. The emphasis would henceforthbe on commercial farmers, not small farmers or peasants. Governmentsubsidies would be much larger for African farmers investing other funds oftheir own. In other words, the larger subsidies would be given to farmers withfunds of their own already. The maximum grant would be a whoppingR100000.

Given that poverty is concentrated in rural areas, land reform needs to beexamined carefully. If output can be increased if small farms practice morelabour-intensive production, then land reform has the potential to promotegrowth and equity at the same time. Unless it takes place on a large scale,

22

Class, Distribution and Redistribution in Post-Apartheid SA

however, it is unlikely to be a source of significant shifts in distribution.There is less poverty in towns and cities. But promoting worker ownershipof firms in manufacturing and services entails a similar transfer of capital toworkers as land reform, except that the capital involved is a factory ormachinery rather than land. If land reform policy is reoriented toward thesemi-privileged, ie prospective medium-sized farmers who can raisesubstantial funds themselves, why should the government not providefinancial assistance to worker-owned firms? Why, we might also ask, dounion investment companies not invest more in worker-owned enterprises?

(d) Welfare ReformThe final pillar in a social democratic agenda must be welfare reform. Publicwelfare is essential for the mitigation of acute poverty, and is especially justin a society where people are willing to work but cannot find work to do. Thepost-apartheid state inherited a welfare system that was remarkablyredistributive in some respects (notably the old-age pension) but profoundlylimited in others (most notably the absence of any support for most of theunemployed). Under Mandela's presidency the only substantial change tothe welfare system was regressive, in that the value of public support forpoor single parents was slashed (ostensibly to allow for higher take-up ratesamong the poor) and the real value of the old age pension declined. Soonafter coming to office, however, the Mbeki government took the bold stepof appointing the 'Taylor' Committee of Inquiry to propose reforms toachieve a comprehensive social security system. The Committee was chargedwith examining not just public welfare systems including unemploymentinsurance), but also public health and the integration of public and privatewelfare and health systems.

One of the ideas prioritised by the Committee was the introduction of abasic income grant (BIG). The introduction of a BIG would make SouthAfrica's welfare regime more coherent and appropriate for a middle-incomelabour-surplus economy. It could also help address the distributionaldilemma of the HPN growth path in that those who gain (employed income-earners) would be taxed extra to pay for the grant. This was the implicit' socialcontract' behind the Scandinavian model - but it only held together whilegrowth was rapid and unemployment relatively low (Nattrass 1999). Butwhether there is sufficient political will in South Africa to shoulder theadditional tax burden remains to be seen.

Samson et al argue that there is room for increased taxation because SouthAfrica's average tax rate is below that of other countries at similar levels of

23

Jeremy Seekings and Nlcoli Nattrass

development (2000:17). The 'People's Budget' (2002) (supported byCOSATU, the South African Council of Churches and the South AfricanNGO Coalition) proposes that part of the needed revenue could be raisedthrough a 'solidarity levy' in the form of a 17.5 per cent surcharge on incometax for the top two quintiles - and the rest in the form of increased taxationof 'the high income group'. This is broadly in line with the COSATU 7th

National Congress Resolution that the cost of the BIG must' fall on the rich'.This of course begs the question as to who comprises the rich. Is it simply

the top decile, or is it the top 40 per cent of income earners? A discoursewhich refers confusingly to 'the rich' and 'the poor' is unlikely to contributeto the kind of social solidarity needed to sustain a BIG over the long-term.By arguing that 'the rich' should absorb all the costs of the BIG, thediscourse of the People's Budget serves to reinforce class antagonismsrather than ameliorate them. As it is, many COSATU workers could findthemselves paying the tax (even if only the proposed 'solidarity levy'), andcould resist it because they had expected it to be levied on the 'rich' - a classthey do not automatically associate themselves as being part of. In its 1998submission to the Jobs Summit COSATU (1998) suggested that a BIG befinanced in part by those earning over R3,000 a month paying back theamount they receive as tax, and those earning over R5,000 per month payingdouble the amount back. A high proportion of COSATU workers (particularlythose in the government sector such as teachers and health workers) wouldfall into the R5,000 per month bracket - and in this regard, the proposal hasclear social democratic aspects to it. However, by 2000, COSATU's positionhad hardened into a discourse that talked only of taxing 'the rich' rather thantaxing income-earners.

In fact, small increases in taxation on union members might be offset byreduced transfers from them to the poor through remittances, if theintroduction of a basic income grant has any crowding out effects onremittances. If the reform of the public welfare system leads to a contractionof the private welfare system, it might be in the interests of trade unioniststo tolerate small increases in taxation (see further Nattrass and Seekings2002).

The proposal for a BIG highlights a central policy dilemma in South Africatoday: how to provide basic income support in a middle-income laboursurplus society. Given that wage earners are protected relative to theunemployed, the moral and political quid pro quo is higher taxation in orderto finance a BIG for the unemployed. The attraction is that the deal does not

24

Class, Distribution and Redistribution in Post-Aparttieid SA

entail any erosion of labour standards or hard-won rights. The downside ishigher taxation, and there is little indication that the organised working classis any more prepared to countenance much higher taxation than it is to agreeto greater labour-market flexibility.

Affirmative action: a noteA social democratic policy agenda can accommodate affirmative action, aslong as this is accurately organised along the lines of disadvantage insociety. Affirmative action should preferably take the form of an investmentin public education which ensures that all children face equal opportunities,regardless of background. Insofar as there are already many people in thelabour market who did not enjoy equality of opportunity, there are strongarguments for affirmative action in the labour market also. But affirmativeaction must be organised according to an appropriate-measure ofdisadvantage.

Despite the dramatic shifts in the determinants of inequality - and theincreasing salience of class - inequality is often reduced to racial inequalityin the minds of many South Africans. As we have noted elsewhere, PresidentMbeki's discourse of 'two nations' treats race as an accurate measure ofdisadvantage (Nattrass and Seekings.2001). The one nation is 'black andpoor', the other 'white and relatively prosperous'. In terms of this world-view, affirmative action to address the legacy of racial discrimination issynonymous with reducing inequality. We have presented a range ofevidence to the effect that inequality is increasingly a function of class,rather than race. During the apartheid era, there was a clear overlap betweenrace and class. However, as Africans moved up the occupational ladder, andas the apartheid system frayed at the edges and then disintegrated, thecontribution of within-group inequality to total inequality rose significantly.In post-apartheid South Africa, inequality is driven by two income gaps:between an increasingly multi-racial middle class and the rest, and betweenthe African urban industrial working class and the African unemployed andmarginalised poor.

Race is thus an ever less adequate proxy for disadvantage. Affirmativeaction which simply ensures that better off African people move into postspreviously occupied by white people might reduce inter-racial inequality butmake no difference to overall inequality, as intra-racial inequality grows.This is what has happened, primarily as a result of the removal ofdiscriminatory state restrictions rather than affirmative interventions, over

25

Jeremy Seekings and Nicoli Nattrass

the past thirty years. If there is to be some redress of persisting inequalitiesof opportunity, then the marker of inequality shouldbe class rather than race.Because class is not an ascriptive characteristic, it has the additionaladvantage over race that it is unlikely to lead to the kinds of division oftengenerated by race-based affirmative action (see also Ramphele 1996, Adam1997).

ConclusionThe imagery of a divided nation is important in highlighting the extreme levelof inequality in post-apartheid South Africa. It is always important to bearin mind the enduring effects of the racialisation of society under apartheid.But Mbeki's portrayal of 'two nations' is an inadequate analysis of SouthAfrican society today. South African inequality is not simply or evenprimarily inter-racial. Declining inter-racial inequality has not reduced overallinequality, and will not do so in future, because the factors that driveinequality have become increasingly significant at the intra-racial level. Ina society that has become dependent on wages and salaries, a reduction ininequality requires a more egalitarian and effective educational system,broader access to employment (through job creation) and reforms to thewelfare system. Insofar as South Africa comprises a divided nation, it is moreaccurate to see it in terms of three broad social groups not two racially-de fined nations: an increasingly multi-racial upper class, comprising not justhigh profile corporate figures but much more broadly the professional,managerial and business classes; a 'middle' group of mostly urban, employedworkers; and a marginalised class of outsiders, comprising many of theunemployed as well as workers in agricultural and domestic employment.

Racial differences in income remain large (due mainly to differences ineducation and the labour market). But while changing the complexion of therich is an important aspect of social transformation in post-apartheid SouthAfrica, it should not be conflated with an egalitarian strategy. A strategywith redistribution at its heart would have to concentrate far more on creatingj obs for the currently unemployed, and on redistributing income through thebudget from rich to poor - perhaps through the introduction of a basicincome grant.

ReferencesAdam, K (1997) 'The politics of redress: South African style affirmative action',

Journal of Modern African Studies 35(2).

Class, Distribution and Redistribution in Post-Apartheid SA

Adler, G and G O'Sullivan (1996) 'Rounding up the usual suspects: recycling thelabour aristocracy thesis', in J Baskin (ed) Against the Current: labour andeconomic policy in South Africa. Johannesburg: Ravan Press.

Adler, G and E Webster (1999) 'Towards a class compromise in South Africa's"Double Transition": bargainedliberalization and the consolidation of democracy',Politics and Society 27(3).

(2000) 'Consolidating democracy in a liberalizing world: trade unions anddemocratization in South Africa', in G Adler and E Webster (eds) Trade Unionsand Democratization in South Africa, 1985-1997. New York: St Martin's Press.

Baskin, J (2000) 'Labour in South Africa's transition to democracy: concertationin a third world setting', in G Adler and E Webster (eds) Trade Unions andDemocratization in South Africa, 1985-1997. New York: St Martin's Press.

BMR (1999) 'Activities, lifestyles and status products of the newly-emergingmiddle class in Gauteng', Research Report 262, Bureau for Market Research,UNISA, Pretoria.

Bhorat, H and J Hodge (1999) 'Decomposing shifts in labour demand in SouthAfrica', South African Journal of Economics 67(3).

Bond, P (2000) Elite Transition: from apartheid to neo-liberalism in South Africa.

London: Pluto Press.Bowles, S and H Gintis (1998) 'Efficient redistribution: new rules for markets,

states and communities', in S Bowles and H Gintis (eds) RecastingEgalitarianism.London and New York: Verso.

Budlender, D (2000) 'Human development', in J May (ed) Poverty and Inequality

in South Africa. Cape Town: David Philip.BusinessMap (1999) Black Economic Empowerment 1999: a moving experience.

Johannesburg: BusinessMap SA.Carter, M and J May (1999) 'Poverty, livelihood and class in rural South Africa',

World Development 27(1).Case, A and A Deaton (1999) 'School inputs and educational outcomes in South

Africa', Quarterky Journal of Economics 458.

Charney, C (1995) 'Voices of a new democracy: African expectations in the newSouth Africa', Research Report 38, Centre for Policy Studies, Johannesburg.

Cichello, P, G Fields and M Leibbrandt (2000) 'Which African workers have gottenahead and by howmuch?: the story of Kwa-ZuluNatal, South Africa: 1993-98'.Unpublished report, School of Economics, University of.Cape Town.

Cortaza, R (1997) 'Chile: the evolution and reform of the labour market', in SEdwards andN Lustig (eds) Labor Markets in Latin America: combining socialprotection with market flexibility. Washington, DC: Brookings Institution Press.

27

Jeremy Seekings and Nicoli Nattrass

COSATU (1998) Submission to the Jobs Summit. Available on www.polity.org.za/govdocs/surnmit/jobsuinmit.html.

Crankshaw, O (1997) Race, Class and the Changing Division of Labour underApartheid. London: Routledge.

Deininger, K and J May (2000) 'Is there scope for growth with equity? The caseof land reform in South Africa'. Working Paper no. 29, Centre for Social andDevelopment Studies, University of Natal, Durban.

Edwards, L (2000) 'Globalisation and the skill bias of occupational employment inSouth Africa'. Unpublished paper, School of Economics, University of CapeTown.

Gelb, S and E Webster (1996) 'Jobs and equity, the social democratic challenge',South African Labour Bulletin 20(3).

Habib, A and V Padayachee (2000) 'Economic policy and power relations in SouthAfrica's transition to democracy', World Development 28(2).

Heintz, J and S Bowles (1996) 'Subsidising employment: wage subsidies and jobcreation', in J Baskin (ed) Against the Current: labour and economic policy inSouth Africa. Johannesburg: Ravan.

Kenny.B andE Webster(1998) 'Eroding the core: flexibility and the resegmentationof the South African labour market', Critical Sociology 24(3).

Lam, D (1999) Generating Extreme Inequality: schooling, earnings and inter-generational transmission of human capital in South Africa and Brazil. ResearchReport # 99-439, Population Studies Center at the Institute for Social Research,University of Michigan.

Maiais,H(1998)SouthAfrica, Limits to Change: thepolitical economy of transition.Cape Town: University of Cape Town Press and London: Zed Books.

( 2001) South Africa, Limits to Change: thepolitical economy of transition.Second edition. Cape Town: University of Cape Town Press and London: ZedBooks.

McGrath, M, J Janisch and C Homer (1997) 'Redistribution through the fiscalsystem in the South African economy'. Paper presented to the EconomicsSociety of South Africa Conference, Potchefstroom.

Michie, J and V Padayachee (1998) "Three years after apartheid: growth, employmentand redistribution', Cambridge Journal of Economics 22(5).

Nattrass, N (1999) 'Globalisation and social accords: a comparative analysis ofSweden, Australia and South Africa', Labour, Capital and Society 32(2).

(2000a) 'Inequality, unemployment and wage-setting institutions in SouthAfrica', Studies in Economics and Econometrics 24(3).

(2000b) "The debate about unemployment in the 1990s', Studies inEconomics and Econometrics 24(3).

28

Class, Distribution and Redistribution in Post-Apartheid SA

(2001) 'High productivity now: a critique of South Africa's growthstrategy', Transformation 45.

Nattrass, N and J Seekings (1997) 'Citizenship and welfare in South Africa:deracialisation and inequality in a labour-surplus economy', Canadian Journalof African Studies 31(3).

(1998a) 'Democratic institutions and development in post-apartheid SouthAfrica', in M Robinson and G White (eds) The Democratic Developmental State:political and institutional design, Oxford Studies in Democratisation. Oxford:Oxford University Press.

(1998b) 'Growth, democracy and expectations in South Africa', in I Abedianand M Biggs (eds) Economic Globalisation and Fiscal Policy. Cape Town:Oxford University Press.

(1998c) 'Changing patterns of inequality in the South African labourmarket', in LPetersson(ed) Post-ApartheidSouthern Africa. London: Routledge.

(2001 a) 'Democracy and distribution in highly unequal economies: the caseof South Africa', Journal of Modern African Studies 39(3).

(2001b) 'Two nations? Race and economic inequality in South Africatoday', Daedalus 130(1).

(2002) 'A basic income grant in South Africa?'. Paper presented at the FirstCongress of the United States Basic Income Guarantee Network, New York(March).

- (forthcoming) FromRace to Class: inequality, unemployment and the socialstructure in South Africa. Cape Town: David Philip.

People's Budget (2002) 'The people's budget'. Available on www.cosatu.org.xa/docs/2001 /peobudO 1 .htm.

Posel, D (1991) The Making of Apartheid, 1948-1961. Oxford: Clarendon Press.RSA (2000a) Budget Review 2000. Department of Finance.

(2000b) Education for All: The South African assessment report. Departmentof Education, Pretoria (March).

Ramphele, M (1996) 'Treading the thorny path to equity', in W James et al (eds)Now that We are Free: coloured communities in a democratic South Africa. CapeTown: IDASA.

Randall, D J (1996)' Prospects for the development of a black business class in SouthAfrica', Journal of Modern African Studies 34(4).

Samson, M, O Babson, K MacQuene, I Van Niekerk and R van Niekerk (2000) "Themacroeconomic implications of poverty-reducing income transfers'. Paper

• presented to the conference: Towards a Sustainable and Comprehensive SocialSecurity System, hosted by the Institute for Social Development, Universityof the Western Cape.

29

Jeremy Seekings and Nicoli Nattrass

Seekings, J (2000) 'Visions of society: peasants workers and the unemployed in achanging South Africa', Studies in Economics and Econometrics 24(3).

(2001) 'Making an informed investment: improving the value of publicexpenditure in primary and secondary schooling in South Africa'. Report for theParliament of South Africa, commissioned by the Standing Committee on PublicAccounts (December).

Simbi, T and M Aliber (2000)' The agricultural employment crisis in South Africa'.Paper presented at the TIPS Policy Forum, Muldersdrift, September.

Simkins, C and I Woolard (2000) 'Analysis of the burden of taxes in the SouthAfrican economy for the years 1995 and 1997'. Report for the Department ofFinance.

South African Health Review (1997).Standing, G, J Sender and J Weeks (1996) The South African Challenge: restructuring

the South African labour market. Geneva: International Labour Organisation.Theron, J (2000) 'Responding to externalisation: Part 1', South African Labour

Bulletin 24(6).Van derBerg, S (2000b) 'An analysis of 1997 school data flowing from work on the

fiscal incidence of social spending in South Africa'. Report for the DepartmentofFinance.

(2001) 'An analysis of fiscal incidence of social spending in South Africa,' 1993-97', Social Dynamics 27( 1).

Webster, E (2001) 'The alliance under stress: governing in a globalizing world', inR Southall (ed) Opposition and Democracy in South Africa. London: Frank Cass.

Whiteford, A and D van Seventer (2000) 'Understanding contemporary householdinequality in South Africa', Studies in Economics and Econometrics 24(3).

30